Asymptotic analysis of the random walk Metropolis algorithm on ridged densities
Authors:
- Alexandros BeskosABAlexandros Beskos
- Gareth RobertsGRGareth Roberts
- Alexandre ThieryATAlexandre Thiery
- Natesh PillaiNPNatesh Pillai
- 1 October 2018
- journal article
- research article
- Published by Institute of Mathematical Statistics in The Annals of Applied Probability
- Vol. 28 (5) , 2966-3001
Abstract
We study the asymptotic behaviour of the Random Walk Metropolis algorithm on “ridged” probability densities where most of the probability mass is distributed along some key directions. Such class of probability measures arise in various applied contexts including for instance Bayesian inverse problems where the posterior measure concentrates on a manifold when the noise variance goes to zero. When the target measure concentrates on a linear manifold, we derive analytically a diffusion limit for the Random Walk Metropolis Markov chain as the scale parameter goes to zero. In contrast to the existing works on scaling limits, our limiting stochastic differential equation does not in general have a constant diffusion coefficient. Our results show that in some cases, the usual practice of adapting the step-size to control the acceptance probability might be sub-optimal as the optimal acceptance probability is zero (in the limit).Keywords
All Related Versions
This publication has 18 references indexed in Scilit:
- Bayesian inverse problems with Gaussian priorsThe Annals of Statistics, 2011
- Riemann Manifold Langevin and Hamiltonian Monte Carlo MethodsJournal of the Royal Statistical Society Series B: Statistical Methodology, 2011
- Bayesian inverse problems for functions and applications to fluid mechanicsInverse Problems, 2009
- Optimal scalings for local Metropolis–Hastings chains on nonproduct targets in high dimensionsThe Annals of Applied Probability, 2009
- Optimal scaling of Metropolis algorithms: Heading toward general target distributionsCanadian Journal of Statistics, 2008
- Weak convergence of Metropolis algorithms for non-i.i.d. target distributionsThe Annals of Applied Probability, 2007
- Optimal scaling of MaLa for nonlinear regressionThe Annals of Applied Probability, 2004
- Optimal scaling for various Metropolis-Hastings algorithmsStatistical Science, 2001
- Weak convergence and optimal scaling of random walk Metropolis algorithmsThe Annals of Applied Probability, 1997
- Monte Carlo Calculations of the Radial Distribution Functions for a Proton?Electron PlasmaAustralian Journal of Physics, 1965